After zypper dup in TW for 2082 packages today, machine does not revive from suspend?

non_space · May 17, 2025, 6:38pm

The bloody comedy of errors continues . . . had to use a blowtorch to get the wrong 5700X out of the socket . . . hammer . . . tongs . . . finally just popped loose . . . and the new 5700G is tack welded into place . . . . : - ))

So, then I was under the impression that the “Wraith” fan that came with the 5700G did not include a tube of thermal paste, as the Cooler Master Hyper 212 fan I bought to get the 5700X going with a fan had included . . . . I got the two fans out for a quick cell shot . . . nothing mentioned on the Wraith box about, “watch out for thermal paste already thinly installed for your convenience,” . . . and I set it down on the table . . . next to the Cooler master tower model. I wanted to ask for opinion on which fan to use . . . and/or preferred thermal paste to use to install either of the fans . . . .

When I picked up the Wraith fan I turned it over and . . . “Hmmm . . . looks like some kind of veneer has been installed that perhaps needs to be removed??? O, no . . . that is the thermal paste already applied, but now has been smeared on one corner. O, I did it again . . . .”

So, when I saw the Wraith had paste I was like, "Avenue of least resistance, the Wraith fan looks smaller than the Cooler Master, but I don’t have thermal paste, so installing the Wraith will move me up a square in this interesting "Yahtze-like game called, “building your own computer for a kick.” But now the question is, having smeared that corner a tad bit, do I still need to get more thermal paste, and then that re-opens the choice on fans . . . ??? OR, dang the torpedoes, there is enough thermal paste on the Wraith to do the job, and the Wraith is a good enough fan that it will be fine to click go on?? And then if I should get more thermal paste, which of the many options is the best choice in pastes to have in use???

hui · May 17, 2025, 6:45pm

The Cooler Master Hyper 212 is the better choice as it is a professional cooler with a 120mm fan (more silent). The wraith which comes with the CPU is a cheap, noisy bare minimum solution.

non_space · May 17, 2025, 7:04pm

@hui

OK, thanks for the kindness of the reply . . . the Cooler Master looks more serious . . . so that is good to know that the Wraith is “cheap and loud” . . . .

Any thermal paste will do, or some are also “cheap and ineffective” and not worth messing with?? The Cooler Master came with “CryoFuze” paste, haven’t searched for it, yet. A couple weeks back I did search for pastes and seems like there are 1000s of options to choose from . . . .

hui · May 17, 2025, 7:07pm

You can use the Cooler Master CryoFuze. It is in the upper price segment of thermal pastes and performs very well in several hardware tests.

karlmistelberger · May 17, 2025, 7:19pm

You may try both out of curiosity. Watch for the thickness of thermal paste. A thin layer works best, e.g.:

The Wraith cooler is still fine. When using the Cooler Master buy a standard paste, e.g.:

Use isopropanol for removal of residues and cleaning, e.g.:

I bought and used both products. Many similar products are available.

non_space · May 17, 2025, 7:36pm

Gents:

Thanks for the replies . . . I did just see the ARTIC MX-4 showing up when I was trying to find out how to contact the seller of the bundle, which seems like Amazon has mixed that data up, connecting me to a seller of stained glass making equipment, instead of the actual computer parts distributor . . . .

As far as the rubbing alcohol goes, seems like the recently installed paste was still soft and largely wiped off with paper towel . . . ??

Since I have the Cooler Master fan already, why waste time with the Wraith?? I don’t like “revisiting” problems that could have been avoided, “measure 5 or 6 times, cut once.”???

karlmistelberger · May 17, 2025, 8:06pm

Residues exist you can’t see, such as thin layers of grease. You may need the cleaner anyway: Disturbing messages from infamous host erlangen - they call it a BUG - #6 by karlmistelberger
If you think its’ a waste of time go with the Cooler Master.

hui · May 17, 2025, 8:26pm

There are several tests available which shows that the Cooler Master outperforms the cheap AMD stock cooler by far. So when you already have both coolers available, it is definitely a waste of time to first use the stock cooler. You will find out that any other cooler is better than the stock cooler (as shown by several hardware tests) and exchange it later against the Cooler Master.

Another thing to think about: why would any manufacturer deliver a good quality cooler for “zero” together with a middle range class CPU? The answer is: they don’t. The included stock cooler is able to keep the CPU cool so that it can operate within specs. But loud and by far not at optimal temperatures for the CPU.

non_space · May 17, 2025, 11:35pm

Yep, these are all valid points. It would be one thing if I would have ordered the right 5700G cpu that included substandard fan for free, but threw in the already properly thinned thermal paste and I just bolted it all up . . . by now I would probably be installing the case components . . . and for my purposes I probably wouldn’t be taxing the cpu very much anyway. And as I am growing older I’m not listening to anybody or any sounds, so if the fan is loud then . . . who is listening to it??

But, I pulled the trigger on the wrong cpu, and then had to order a fan, and in looking at it, it seems to have tubing to carry the heat away from the cpu and then blow it sideways out the back, whereas the Wraith doesn’t have tubes, just fins right off the cpu and the fan is either pulling the heat up into the cae or blowing it back down onto the cpu . . . ??

At this juncture it makes sense to get another tube of thermal paste, possibly yes, clean the Hyper212 base plate with alcohol to clean up vagrant greases, as there was a plastic shield there that was removed before the first installation . . . and then play it again.

In some sense it was due to Karls advice that I got the external power switch to test out the mobo components before installing into the case . . . which saved me having to mess around with the items in the case or pull them back out, etc.

In any project there has to be some “momentum” established, and out of that movement, the project moves, but included in movement can be . . . pilot error. A few more days to get the situation back on track . . . and then on to installing the parts into the case.

karlmistelberger · May 18, 2025, 5:19am

non_space · May 19, 2025, 4:47pm

Gearing up to run the maiden install of TW on the 1TB drive in a bit. The intention is to replicate the multi-boot of 6 or 7 distros that I have now on the '12 Mac Pro in the new FrankenB550Ryzen7GBerry machine.

One or two bootloaders for back up purposes . . . figuring to over time cut a number of 50GB partitions for the / filesystems, and then either one or two larger partitions for the various /home directories.

Undecided on the proper sizing for the swap partition, as some distros in the ubuntu flavorings do not use a swap partition any more, they use a swap file instead. Does TW require a swap partition? I still use them on the Mac Pro, in the old “1 - 1.5x” of RAM . . . have 16 GB on the Mac Pro, now bouncing to 32GB on the new machine. 1TB is enough for my needs, but do I need to cut out 32 GB of space for the swap, or would say 16GB be enough??

Not a gamer, sometimes I run Distributed.net stats, but mostly checking forums for answers to questions like this one . . . don’t recall a “swap event” happening over the years . . . . Is there a prudent choice to make in the swap designation of space???

karlmistelberger · May 19, 2025, 7:44pm

Instead of replicating the systems you may try and move the drive(s).

I abandoned swap years ago and never had any problems. You may try KISS:

erlangen:~ # fdl /dev/nvme1n1
Disk /dev/nvme1n1: 1.82 TiB, 2000398934016 bytes, 3907029168 sectors
Disk model: Samsung SSD 990 EVO 2TB                 
Disklabel type: gpt

Device          Size Type
/dev/nvme1n1p1  100M EFI System
/dev/nvme1n1p2  1.8T Linux filesystem
erlangen:~ #

Tumbleweed needs no swap and caches files (19Gi), which makes Infamous Host Erlangen even faster:

erlangen:~ # free -h
               total        used        free      shared  buff/cache   available
Mem:            27Gi       8.0Gi       806Mi       242Mi        19Gi        19Gi
Swap:             0B          0B          0B
erlangen:~ #

non_space · May 19, 2025, 9:46pm

Thanks for the reply on the swap question . . . similar to KISS I also like, “less is more” as a mantra of supreme guidance.

As far as “moving the drives” goes . . . that might work, in that case I have distros spread across a few drives, . . . along with some OSX installs, which likely would not work too well?? Since the Mac Pro is in fact still running OK, I hadn’t thought about stripping it for the drives . … but in starting fresh with the new platform . . . freshly.

I like “variety” . . . “jack of all trades,” rather than master of one . . . I’ll have to see how it goes. Obviously with the “rolling hardware” concept as the motivation, drives could be added in, and or removed, over time . . . .

malcolmlewis · May 19, 2025, 9:55pm

@non_space just run zram… no swap

non_space · May 19, 2025, 11:25pm

Malcolm:

Thanks for that suggestion . . . didn’t know what that was or is, ran a google on it, found an article on fosspost going through the steps needed to change over to it . . . .

But, then the article concluded with:

If you have a large amount of RAM on your system (32 GB or more), then perhaps you don’t need to use zRAM, and it may affect your performance negatively (because of the compression and decompression time), but other than that, it should provide you with better performance because swapping (which your system will do often on smaller amounts of RAM) to RAM is faster than swapping to disk.

karlmistelberger · May 20, 2025, 5:50am

The author of Enable Zram On Linux For Better System Performance reiterates in a comment to the article:

"We mentioned the reason swiftly in the article: If you have a large amount of RAM (simply more than your needs), then you are basically using zRAM for nothing, while also wasting some of your CPU power for compressing and and decompressing your RAM contents.

zRAM is not free, it comes with the cost of using some CPU power to deal with the data coming in and out of the zRAM disk created on your system memory. If you already have the RAM you need and you don’t need more, then you are basically wasting CPU power for nothing and creating an additional headache for nothing. Your system already can handle RAM usage in its current setup.

While CPU time for compression and decompression is generally negligble, there is no need to lose it if you have a system with large RAM that you don’t need. After all, the point of zRAM is to provide more RAM to your system on the cost of little CPU ticks."

Using zRAM adds complexity to a system designed for utmost simplicity. You may want to investigate behaviour of the basic system first.

Fedora 42 uses zRAM by default. I booted into Live Fedora KDE Plasma Desktop 42 and installed to a 250 GB Samsung 850. Automatic partitioning is:

root@fedora:~# fdl /dev/sda
Disk /dev/sda: 232,89 GiB, 250059350016 bytes, 488397168 sectors
Disk model: Samsung SSD 850 
Disklabel type: gpt

Device       Size Type
/dev/sda1    600M EFI System
/dev/sda2      1G Linux extended boot
/dev/sda3  231,3G Linux filesystem
root@fedora:~#

File systems are:

root@fedora:~# lsb /dev/sda
PATH      LABEL  UUID                                 FSTYPE MOUNTPOINT
/dev/sda                                                     
/dev/sda1        C697-E9A7                            vfat   /boot/efi
/dev/sda2        05ce8ff8-28d2-4ff9-a820-1e265957fa4e ext4   /boot
/dev/sda3 fedora ffcaea0e-8571-43ff-a030-416b82220b26 btrfs  /home
root@fedora:~#

Memory usage is:

root@fedora:~# free -h
               total        used        free      shared  buff/cache   available
Mem:            31Gi       1,2Gi        29Gi       229Mi       1,0Gi        29Gi
Swap:          8,0Gi          0B       8,0Gi
root@fedora:~#

I move drives between different chassis e.g. when troubleshooting. This ensures coherent and consistent maintenance:

root@fedora:~# inxi -zSMCm
System:
  Kernel: 6.14.5-300.fc42.x86_64 arch: x86_64 bits: 64
  Console: pty pts/0 Distro: Fedora Linux 42 (KDE Plasma Desktop Edition)
Machine:
  Type: Desktop Mobo: ASRock model: Z170 Pro4S serial: <filter> UEFI: American Megatrends v: P7.50
    date: 01/23/2018
Memory:
  System RAM: total: 32 GiB available: 31.03 GiB used: 1.45 GiB (4.7%) igpu: 256 MiB
  Array-1: capacity: 64 GiB slots: 4 modules: 2 EC: None
  Device-1: ChannelA-DIMM0 type: no module installed
  Device-2: ChannelA-DIMM1 type: DDR4 size: 16 GiB speed: 2133 MT/s
  Device-3: ChannelB-DIMM0 type: no module installed
  Device-4: ChannelB-DIMM1 type: DDR4 size: 16 GiB speed: 2133 MT/s
CPU:
  Info: quad core model: Intel Core i7-6700K bits: 64 type: MT MCP cache: L2: 1024 KiB
  Speed (MHz): avg: 800 min/max: 800/4200 cores: 1: 800 2: 800 3: 800 4: 800 5: 800 6: 800
    7: 800 8: 800
root@fedora:~#

While Fedora currently sits in the chassis of the venerable 6700K it will run on the 5700G without making any modifications.

non_space · May 20, 2025, 2:12pm

Thanks for the clarifications on the zRAM . . . I think for me 32GB of RAM exceeds my daily needs by quite a bit . . . simple is indeed good. Looks like if and when we arrive at the time in which 32GB is like today’s 4GB, then the zRAM approach could be used.

I looked at Fedora sometime last year when I was looking for a 7th distro for the 7th day, and the live system booted up fine, but the installer only knows how to work with the whole drive, I couldn’t use “custom” and direct it into a specific partition. So your idea of putting it on its own 250GB drive to play on its own is the Way, the installer did not appear to play well with others . . . .

non_space · May 21, 2025, 3:08am

Happy to post this data . . . have yet to check the resume from suspend function. : - )))

~> inxi -zSMCm
System:
  Kernel: 6.14.6-1-default arch: x86_64 bits: 64
  Desktop: MATE v: 1.28.2 Distro: openSUSE Tumbleweed 20250515
Machine:
  Type: Desktop System: Micro-Star product: MS-7C56 v: 6.0
    serial: <superuser required>
  Mobo: Micro-Star model: B550-A PRO (MS-7C56) v: 2.0
    serial: <superuser required> UEFI: American Megatrends LLC. v: A.H0
    date: 03/11/2024
Memory:
  System RAM: total: 32 GiB available: 27.3 GiB used: 2.29 GiB (8.4%)
  Array-1: capacity: 128 GiB slots: 4 modules: 2 EC: None
  Device-1: Channel-A DIMM 0 type: no module installed
  Device-2: Channel-A DIMM 1 type: DDR4 size: 16 GiB speed: 2133 MT/s
  Device-3: Channel-B DIMM 0 type: no module installed
  Device-4: Channel-B DIMM 1 type: DDR4 size: 16 GiB speed: 2133 MT/s
CPU:
  Info: 8-core model: AMD Ryzen 7 5700G with Radeon Graphics bits: 64
    type: MT MCP cache: L2: 4 MiB
  Speed (MHz): avg: 2995 min/max: 400/4673 cores: 1: 2995 2: 2995 3: 2995
    4: 2995 5: 2995 6: 2995 7: 2995 8: 2995 9: 2995 10: 2995 11: 2995 12: 2995
    13: 2995 14: 2995 15: 2995 16: 2995

non_space · May 21, 2025, 4:05am

Hmmm . . . no resume from suspend . . . likely because there is no swap to suspend to??

Hopefully I can add a swap partition via GParted . . . and then add it in to the fstab data???

So back to the, how much swap is the right amount of swap so that the machine will resume from suspend via keystroke??

Or is this relating to having an ext power switch?? I have to click on the power switch to get to the log in window, rather than mouse click or keyboard click . . . .

mrmazda · May 21, 2025, 4:08am

Looks like there have been 5 BIOS updates since then.